Skip to content

Latest commit

 

History

History
122 lines (95 loc) · 3.42 KB

README.md

File metadata and controls

122 lines (95 loc) · 3.42 KB

Remote Code Execution Engine

Go

This repository contains a remote code execution engine that allows users to submit code in various programming languages, execute it in isolated Docker containers, and retrieve the output. The engine supports both x86_64 and arm64 architectures.

Features

  • Supports C++ and Go programming languages.
  • Executes code in isolated Docker containers.
  • Supports both x86_64 and arm64 architecture machines.
  • Cleans up zombie containers to avoid memory leaks.
  • Provides a REST API for code submission and execution.
  • Restricts the usage of system resources (Memory, CPU, max processes, max files, max file size)
  • Kills a container if it is taking more than a minute to complete the execution.
  • Supports custom docker images and compilation commands for each programming language.

Prerequisites

  • Docker
  • Go

Setup

  1. Clone the repository:

    git clone https://github.com/srujangit123/Remote-Code-Execution-Engine.git
cd Remote-Code-Execution-Engine

  2. Build the Docker images: This builds the default docker images and uses the default config (config.yml) Skip this step if you want to use custom docker images and commands. Make sure you modify the config.yml if you are using custom docker images and commands

    bash scripts/build_docker.sh

  3. Build the server:

    go build -o server cmd/server.go

Configuration

The configuration file config.yml specifies the settings for each supported language. Here is an example configuration:

cpp:
  extension: ".cpp"
  image: "cpp_arm64:latest"
  command: "/usr/bin/run-code.sh {{LANGUAGE}} {{FILE}} {{INPUT}}"
golang:
  extension: ".go"
  image: "golang_arm64:latest"
  command: "/usr/bin/run-code.sh {{LANGUAGE}} {{FILE}} {{INPUT}}"

Variables available in the command config

  • {{LANGUAGE}} - programming language
  • {{FILE}} - Code file with the extension as specified in the config
  • {{INPUT}} - Input file if the input is provided by the user

These variables are replaced with appropriate values before creating the code container.

Running the server

To start the server, run the following command:

./server

Flags

  • --code-dir The default directory where the code files will be stored is /tmp/ A separate directory is created for every language. To change the directory where the code files will be stored(eventually removed by the garbage collector), use
./server --code-dir /path/to/code/dir
  • --resource-constraints By default, resource constraints are turned off to improve the performance, if you want to enable it, use
./server --resource-constraints true

API

Supported Languages

  • URL: /api/v1/languages
  • Method: GET

Submit Code

  • URL: /api/v1/submit
  • Method: POST
  • Content-Type: application/json
  • Request Body:
{
    "code": "base64_encoded_code",
    "input": "base64_encoded_input",
    "language": "cpp"
}
  • Response:
{
    "output": "execution_output"
}

Example

To submit a code execution request, you can use the following curl command:

curl -X POST http://localhost:9000/api/v1/submit \
     -H "Content-Type: application/json" \
     -d '{"code": "base64_encoded_code", "input": "base64_encoded_input", "language": "cpp"}'